In over-the-counter (OTC) markets for assets such as currencies, derivatives, and commodities, staggering sums of money are tied to single, critical financial benchmarks.  The London Interbank Offered Rate (LIBOR), for example – often referred to as “the world’s most important number” – determines interest payments on instruments ranging from student loans and mortgages to synthetic derivatives across the globe.  In 2016, estimates of notional exposure to U.S. dollar LIBOR totaled about $200 trillion – 10 times U.S. GDP that year.  The WM/Reuters foreign exchange benchmark is another example.  Its impact extends to retirement funds and stock markets, where pension funds and stock indices (such as the S&P 500 and Dow Jones Industrial Average) all reference WM/Reuters rates in valuing investments and stocks denominated in foreign currency.  Correspondingly, minuscule variations in a benchmark’s value affect vast amounts of assets and transactions for hundreds of millions of people.

In my article, Benchmark Competition, I explore how such benchmarks have grown so dominant and why market participants continue to treat them as “one size fits all,” even when they are suboptimal.

The ubiquity of these benchmarks is not difficult to understand: They have substantially harmonized otherwise decentralized, opaque dealer markets (where a dealer such as Goldman Sachs or other large market-making entity is a counterparty to each transaction).  By coordinating the price discovery-like functions of dealers, benchmarks have enhanced price transparency, lowered the cost of contracting, expanded the range of hedging opportunities, and generated enormous liquidity through network effects.  Upon initial observation, this narrative fits the prevailing view of financial regulation: Because sophisticated market participants, through wealth-maximizing behavior, tend to select structures that maximize efficiency, aggregate social welfare is maximized, meaning that observed equilibria are likely the most efficient equilibria.

I argue that this understanding is incomplete.  Significant yet overlooked distortions have resulted from treating benchmark rates as “one size fits all,” distortions that highlight a fundamental misalignment between what is privately optimal and what is socially optimal in OTC markets.

First, the dominance of certain benchmarks has helped entrench the natural oligopoly of dealers in OTC markets.  The most powerful dealers, already concentrated, have outsize influence over the price discovery functions of benchmarks.  As benchmarks become critical to the functioning of the economy, those dealers can correspondingly gain some control over legal and regulatory aspects of the market as well.  LIBOR has become systemically important in the same way that a bank or financial institution can be, which has already reduced the credibility of regulatory discipline.  In 2008, regulators at the Bank of England seem to have implicitly endorsed LIBOR manipulation to avoid further undermining global confidence in the banks.

Second, the widespread hardwiring of benchmarks throughout transactions has made manipulating them particularly tempting.  A small tweak to a benchmark will have a disproportionately large financial effect.  For example, economists have shown that in a single quarter, a 25 basis point (0.25 percent) change in LIBOR could net $337 million in interest revenue for JPMorgan, and $936 million for Citigroup, based on their outstanding exposures.  And as both the entities involved in the benchmark-setting process, as well as the counterparties to transactions that reference those benchmarks, dealers are uniquely well positioned to manipulate.

Finally, network effects and path dependencies can promote inefficient pooling around the default benchmark, even if it is suboptimal.  Market participants may choose to transact at a benchmark simply because everyone else is doing so.  Entire ecosystems of additional products can sprout up around a particularly successful benchmark.  The switching costs become enormous.  As network effects proliferate and the same benchmark is used again and again, the likelihood of competition in the form of other benchmarks or referents becomes vanishingly small.  In this environment, not only will a dominant benchmark provider lack adequate incentive to promulgate additional benchmarks, it might also fail to invest in improvements to the existing benchmark or adequately monitor for wrongdoing.

I argue that, so long as a benchmark remains entrenched in an ecosystem dominated by powerful institutions, there is little likelihood of innovation or competition from socially-beneficial alternative benchmarks, and market-based discipline will remain ineffective.  Instead, suboptimal yet systemically important benchmarks will persist, with skyrocketing switching costs for market participants and zero incentives to develop better benchmarks.

One size cannot and should not fit all.  An alternative competitive equilibrium should be considered – one where multiple benchmarks compete.

Existing reform proposals emphasize calculation-methodology changes, enforcement, or a government-created benchmark.  All overwhelmingly assume that a single benchmark will continue to dominate.  That ignores the deeper pathologies identified in this article and the problems of one-size-fits-all benchmarks.

Calculation reform, while useful, comes at a cost.  For example, widening the sampling window for the WM/Reuters benchmark calculation has introduced tracking error and potentially lowered the utility of such a rate to end users.  Moreover, total immunity to manipulation is impossible, and attempting to achieve it would be exceedingly costly.

Enforcement and compliance will no doubt remain important.  However, the patchwork of fraud, manipulation, and antitrust regimes, and the difficulty of measuring harm or disgorging profits pose significant obstacles to both ex ante deterrence and ex post discipline.  Nor do such proposals address the problem of regulatory capture due to the systemic importance of a benchmark, as occurred with LIBOR in 2008.

Relying on the government as a benchmark provider is also problematic, and likely to be slow, cumbersome, and costly.  There is also no guarantee that the government would get it right.  In the U.S., regulators have settled on the secured overnight financing rate (SOFR), a measurement of banks’ overnight borrowing rate (secured by Treasury securities), as the sole designated LIBOR replacement.  LIBOR had flaws, to be sure.  But so does SOFR: The markets for SOFR can be finicky and overly dependent on volatility spurts related to funding market idiosyncrasies.  SOFR, an average of past transactions, should also not be viewed as a satisfactory substitute for LIBOR, a credit-sensitive term rate.  Importantly, SOFR is backward-looking, while LIBOR is forward-looking.  For many borrowers, interest rates pegged to future economic movements will be much more useful than those that will always lag the market.

These differences mean that, to fully transition to SOFR, market participants will need to understand and calculate mathematical relationships between LIBOR and SOFR for many contracts with payment obligations extending far into the future, an exceedingly costly endeavor.  Any issues will be exacerbated because banks, the main players in OTC markets, are better off when they can match their lending revenue with their borrowing costs.  LIBOR, which measures banks’ cost of borrowing from each other, was actually an excellent rate at which to lend: Banks’ revenue from loans made at LIBOR would match their cost of borrowing from each other, allowing a match between assets and liabilities.

Competition among multiple benchmarks, by contrast, can benefit participants market-wide by encouraging innovation, transparency, the development of better information, and entry by more efficient providers. Benchmarks would be nimbler, less systemically important, and less tempting and more difficult to manipulate.

Additional benchmarks would, critically, significantly lower the cost for end users of switching from one benchmark to another.  Stagnation around a single benchmark could be more easily avoided, and wealth-promoting innovation and updating around existing and new benchmarks could occur (with the proper incentives).  Reduced hardwiring of a single benchmark would also make market discipline more effective by reducing not only the temptation to manipulate (by lowering the upside), but also a benchmark’s systemic impact.  As an analogy, the S&P 500, the Dow Jones, NASDAQ, and Russell 2000 provide varying but similar measures of the stock markets’ performance.  If the Dow Jones were compromised, alternatives would make opting out of referencing it more realistic and comparatively much less costly.

A one-size-fits-all benchmark is not a realistic goal – and staying within that paradigm will likely only replicate existing problems and result in unanticipated and unnecessary cost and risk to the financial system.  As the president of the Federal Reserve Bank of New York stated in a speech in September 2019, “[c]ontracts that reference U.S. dollar LIBOR continue to be written, which only serves to increase the level of systemic risk.”

This post comes to us from Sue S. Guan, the post-doctoral research scholar in the Program in the Law & Economics of Capital Markets at Columbia Law School and Columbia Business School. It is based on her recent article, “Benchmark Competition,” available here.